Yes, do you not need an output capacitor? An active load with the same cathode resistor as the lower tube (or both bypassed) would have a gain of mu/2, or about x50.
There would be some degree of distortion cancellation using the active load, if the impedance of the driven circuit is high, or if a buffer is used.
The output impedance without a buffer will be about half as much if the cathode resistors are both bypassed.
Last edited:
I do not need (in fact, cannot use) an output capacitor — the circuit is providing both the polarizing voltage and drive signal for a primitive electrostatic panel (I have no way of estimating the load impedance, other than to guess "high"). Gain sounds about right. Component count is attractive. I can size the bypass capacitor to shape the tone. I understand what the circuit is doing. I think I'm going to give this one a try tonight!
Ok, just use the exact same values of cathode resistor and bypass capacitor on both halves.
Both halves of the triode. Don't run the top triode without a cathode resistor, bypassed or not.
I suspect it will be less linear since the bottom triodes anode load is close to ra/mu due the the CF action of the top tube. If you want good linearity then try a mu follower. You can obtain exceptionally low distortion in a 12AX7 using that topology.
Cheers
Ian
Won't it just follow the 0v grid line in sympathy with the bottom section?
Sorry for all the edits.
Identical cathode resistors and bypass capacitors would center the bias point, which is what I'm trying to avoid. I realize it's an uncommon application.
Last edited:
I suppose you could get a depletion mode FET to work. You have exactly the same set of biasing issues that occur in setting up a DC coupled cathode follower. Remember that reverse bias is necessary, when depletion mode devices, including tubes/valves, are employed.
The beauty of enhancement mode FETs, in the voltage follower role, is the great ease in setting things up. The requisite forward bias occurs automatically, when DC coupling is used. You know the gate voltage and the drain voltage. As the source follows the gate, pick a drain current and use Ohm's Law to calculate the load resistance value. Yes, assuming the internal resistance of the FET is 0 Ω is wrong, but the approximation is good enough.
I will add a word of warning. Experience has shown that a protective Zener diode should be installed between gate and source. Because the FET conducts immediately and the tube delays conduction, when power is applied, the operating conditions calculated do not obtain. The Zener diode ensures that the gate to source potential limit does not get exceeded by start up transients.
Depletion source follower
Many thank for replying Eli,
I'm not sure it will be so difficult. I successfully modified a K&K cascode depletion fet CCS to be a cascode deletion fet source follower. I feed the signal into the gate and take the output from the source. The upper cascode is biased in a similar way to the CCS. I added back to back protection zeners although the K&K CCS doesn't have them.
I use it to drive the EL519 in EMT (A2) in my venerable diy EAR 859 amp. I think it sounds good. I think it sounds "better" than the PCC88 cathode follower; though a large gate stopper and a source "stopper" resister are required.
I don't know much so maybe I was just lucky that it worked.
best regards
tim
Many thank for replying Eli,
I'm not sure it will be so difficult. I successfully modified a K&K cascode depletion fet CCS to be a cascode deletion fet source follower. I feed the signal into the gate and take the output from the source. The upper cascode is biased in a similar way to the CCS. I added back to back protection zeners although the K&K CCS doesn't have them.
I use it to drive the EL519 in EMT (A2) in my venerable diy EAR 859 amp. I think it sounds good. I think it sounds "better" than the PCC88 cathode follower; though a large gate stopper and a source "stopper" resister are required.
I don't know much so maybe I was just lucky that it worked.
best regards
tim
You have drawn a linear load line. OK for a resistor, but a triode is not a resistor.SlunkBoy said:
To get good linearity from a lightly-loaded SRPP you need the upper and lower parts to be identical, not significantly different - sometimes known as the 'half-mu' circuit for obvious reasons. By 'lightly-loaded' I mean the load must be little more than a valve grid plus a high value grid leak resistor, with not too much Miller capacitance. Then the non-linearities of the two anode impedances cancel. Provided the valve has fairly constant mu (e.g. 12AX7/ECC83) you get low distortion. The alternative is to use a mu-follower, which achieves low distortion in a different way: high impedance rather than balancing impedance.
A non-degenerated triode valve typically has a lower AC impedance than a linear resistor which matches the DC conditions, and hence is worse as an anode load unless it happens to match the amplifying valve (which would normally require balanced operation i.e. same bias for both).
I was able to play around with some simple active-load circuits last night. I first put together a SRPP with matched load resistors. I didn't connect a source or a load as I was only interested in making sure the circuit performed as expected for DC conditions (I had also never experimented with floating the bias point before). In that regard, it was a success. I then reconfigured the circuit — removing the upper triode's cathode resistor and setting the bias point at 250v. The lower valve's cathode resistor must've been a little off, because it biased to 220v — but it otherwise performed as expected. I then added a source, load, and bypass capacitor for the lower triode's cathode resistor. As everyone here expected, the results were unimpressive. The gain was far too low and the distortion was certainly not an improvement over the plate load resistor circuit. That said, it was a good learning experience and I was pleased that my calculations generally matched my real world results.
I'm going to try a variation of the first schematic I posted tonight. I think I've worked the values out such that I can get rid of the capacitor in the feedback chain. My understanding is that the ratio of the feedback resistor and the cathode resistor in the first stage simply needs to be selected such that the input triode has the correct voltage dropped across it during quiescent conditions. I'll report back with my results.
I'm going to try a variation of the first schematic I posted tonight. I think I've worked the values out such that I can get rid of the capacitor in the feedback chain. My understanding is that the ratio of the feedback resistor and the cathode resistor in the first stage simply needs to be selected such that the input triode has the correct voltage dropped across it during quiescent conditions. I'll report back with my results.
- Status
- Not open for further replies.